{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/its-moving-a-probabilistic-model-for-causal","title":"It's Moving! A Probabilistic Model for Causal Motion Segmentation in Moving Camera Videos","arxiv_id":"1604.00136","date":"2016-04-01","proceeding":null,"authors":["Pia Bideau","Erik Learned-Miller"],"abstract":"The human ability to detect and segment moving objects works in the presence\nof multiple objects, complex background geometry, motion of the observer, and\neven camouflage. In addition to all of this, the ability to detect motion is\nnearly instantaneous. While there has been much recent progress in motion\nsegmentation, it still appears we are far from human capabilities. In this\nwork, we derive from first principles a new likelihood function for assessing\nthe probability of an optical flow vector given the 3D motion direction of an\nobject. This likelihood uses a novel combination of the angle and magnitude of\nthe optical flow to maximize the information about the true motions of objects.\nUsing this new likelihood and several innovations in initialization, we develop\na motion segmentation algorithm that beats current state-of-the-art methods by\na large margin. We compare to five state-of-the-art methods on two established\nbenchmarks, and a third new data set of camouflaged animals, which we introduce\nto push motion segmentation to the next level.","url_abs":"http://arxiv.org/abs/1604.00136v1","url_pdf":"http://arxiv.org/pdf/1604.00136v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"motion-segmentation","task_name":"Motion Segmentation"},{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[],"datasets_introduced":[{"slug":"camouflaged-animal-dataset","name":"Camouflaged Animal Dataset","full_name":"Camouflaged Animal Dataset"}],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1604.00136","atlas_url":"https://app.syntology.ai/?focus=1604.00136","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}